X-Git-Url: https://www.fleuret.org/cgi-bin/gitweb/gitweb.cgi?p=dyncnn.git;a=blobdiff_plain;f=README.txt;h=85cf8ba1e9373a3d595dd1106b4c9021731240f2;hp=8a832509880cfdb94b32cab0042dd6da9a1c0879;hb=a79fda5dc501909019e40b185760be3fbaa4d12d;hpb=1ac95d100c59265ae76df4fea676f016e1b19a50

diff --git a/README.txt b/README.txt
index 8a83250..85cf8ba 100644
--- a/README.txt
+++ b/README.txt
@@ -1,17 +1,31 @@
 
 This is an implementation of a deep residual network for predicting
-the dynamics of 2D shapes.
+the dynamics of 2D shapes as described in
 
-This package is composed of two main parts: A simple 2d physics
-simulator called 'flatland' written in C++, to generate the data-set,
-and a deep residual network 'dyncnn' written in the Lua/Torch7
-framework.
+  F. Fleuret. Predicting the dynamics of 2d objects with a deep
+  residual network. CoRR, abs/1610.04032, 2016.
+
+  https://arxiv.org/pdf/1610.04032v1
+
+This package is composed of a simple 2d physics simulator called
+'flatland' written in C++, to generate the data-set, and a deep
+residual network 'dyncnn' written in the Lua/Torch7 framework.
 
 You can run the reference experiment by executing the run.sh shell
-script. It will generate the data-set of 50k triplets of images, train
-the deep network, and output validation results every 100 epochs.
+script.
+
+It will
+
+  (1) generate the data-set of 50k triplets of images,
+
+  (2) train the deep network, and output validation results every 100
+      epochs. This take ~30h on a GTX 1080.
+
+  (3) generate two pictures of the internal activations.
+
+  (4) generate a graph with the loss curves if gnuplot is installed.
 
 --
 Francois Fleuret
-Oct 7, 2016
+Oct 21, 2016
 Martigny